Exogenous PPAR 2 but Not PPAR 1 Reactivates Adipogenesis
Total Page:16
File Type:pdf, Size:1020Kb
Downloaded from genesdev.cshlp.org on September 23, 2021 - Published by Cold Spring Harbor Laboratory Press RESEARCH COMMUNICATION ␥ differential isoform function. Rationally engineered PPAR knockdown by transcription factors potentially provide a powerful tool engineered transcription for targeted regulation of endogenous genes by combin- ing a functional transcription regulatory domain with a factors: exogenous PPAR␥2 customized DNA binding domain that can bind to a spe- but not PPAR␥1 cific sequence within the target gene. C2H2 zinc finger proteins (ZFPs) can be engineered to bind with high reactivates adipogenesis specificity to wide a diversity of DNA sequences (Des- jarlais and Berg 1992; Choo and Klug 1994; Jamieson et Delin Ren,1 Trevor N. Collingwood,2 al. 1994; Rebar and Pabo 1994; Greisman and Pabo 1997). Edward J. Rebar,2 Alan P. Wolffe,2 Previous studies have demonstrated the utility of both and Heidi S. Camp1,3 engineered activator- and repressor-ZFPs in the regula- tion of endogenous chromosomal loci (Bartsevich and 1Department of Molecular Science and Technology, Pfizer Juliano 2000; Beerli et al. 2000; Zhang et al. 2000; Liu et Global and Research Development, Ann Arbor, Michigan al. 2001). Our goal for this study was to selectively in- 48105, USA; 2Sangamo BioSciences Inc., Pt. Richmond, hibit expression of the PPAR␥2 isoform in the adipo- California 94804, USA genic mouse 3T3-L1 cell line by utilizing engineered zinc finger repressor proteins. To determine functional differences between the two splice variants of PPAR␥ (␥1 and ␥2), we sought to se- Results and Discussion ␥ lectively repress 2 expression by targeting engineered The mouse PPAR␥ gene spans >105 kb (Zhu et al. 1995). ␥ zinc finger repressor proteins (ZFPs) to the 2-specific Coding exons 1 to 6 are conserved between the ␥1 and ␥2 promoter, P2. In 3T3-L1 cells, expression of ZFP55 re- isoforms (Fig. 1A) and transcription of these is driven by sulted in >50% reduction in ␥2 expression but had no an upstream promoter (P1) that also drives expression of effect on ␥1, whereas adipogenesis was similarly reduced two untranslated ␥1-specific exons, A1 and A2. The ad- by 50%. However, ZFP54 virtually abolished both ␥2 ditional amino acids at the amino terminus of ␥2 are and ␥1 expression, and completely blocked adipogenesis. encoded by an additional exon, B1, that is uniquely regu- Overexpression of exogenous ␥2 in the ZFP54-expressing lated by a separate promoter (P2) lies >63 kb downstream ␥ cells completely restored adipogenesis, whereas overex- of P1. DNase I digestion of the endogenous PPAR gene locus in 3T3-L1 cells revealed two DNase I hypersensi- pression of ␥1 had no effect. This finding clearly identi- ␥ tive sites in the vicinity of the proximal P2 promoter fies a unique role for the PPAR 2 isoform. that represent regions of accessible chromatin at an en- Received October 12, 2001; revised version accepted dogenous locus (Fig. 1B; DHS1 and DHS2). Two six-fin- November 15, 2001. ger ZFPs (ZFP54 and ZFP55; Fig. 1C) linked to the KRAB transcriptional repression domain (Margolin et al. 1994) were designed to bind specifically to 18-bp sequences The nuclear hormone receptor PPAR␥ is essential for within the DHS1 shown in Figure 1B. Each ZFP bound cellular differentiation and lipid accumulation during its cognate site on naked DNA with high affinity adipogenesis (Barak et al. 1999; Kubota et al. 1999; Rosen (Kd = 20 and 44 pM, respectively). et al. 1999). The adipocyte-specific ␥2 isoform differs The ZFP54 and ZFP55 repressor proteins were ex- from the more widely expressed ␥1 in that it contains pressed retrovirally in 3T3-L1 cells to similar levels (Fig. additionally 30 amino acid residues at the amino termi- 2A). Nontransduced wild-type cells and stable pools of nus (Kliewer et al. 1994; Tontonoz et al. 1994a; Zhu et al. infected cells expressing each ZFP, as well as control 1995). Evidence suggests these residues contribute to a cells retrovirally transduced to express LacZ, were in- constitutive transcription activation function that is duced to differentiate and initiate the adipogenic path- 5–10-fold greater than in ␥1 (Werman et al. 1997). way. Total PPAR␥ mRNA level was determined at 0, 2, PPAR␥2 is selectively expressed in adipose tissue (Fajas and 5 d post induction. Over the 5-d time course PPAR␥ et al. 1997) and is strongly up-regulated during adipogen- expression increased ∼5.8- to 6.0-fold in both the wild- esis (Tontonoz et al. 1994b; Wu et al. 1998), suggesting a type and LacZ control cells (Fig. 2B). Up-regulation of specific role for this isoform in fat cell differentiation. total PPAR␥ expression was reduced slightly in the pres- Nevertheless, a specific role for ␥2 that could not be ence of ZFP55 (4.8-fold compared to 6.0-fold in wild type) substituted by ␥1 has not been clearly determined. but completely inhibited by ZFP54. However, a time The ability to selectively knock out or knock down course analysis of expression of the individual PPAR␥ the expression of a specific gene provides a powerful ap- isoforms revealed that although ZFP54 effectively inhib- proach for understanding its biological function. The tar- ited expression of both isoforms, ZFP55 selectively re- geting of individual mRNA splice variants offers an even pressed PPAR␥2by∼50% and had little effect on greater level of selective control and understanding of PPAR␥1 compared with wild-type cells (Fig. 2C). An- other ZFP–KRAB fusion protein that binds to a sequence that is absent in the PPAR␥ gene has no effect on PPAR␥ [Key Words: PPAR␥1; PPAR␥2; adipogenesis; zinc finger protein] expression or adipogenesis (data not shown). The effects 3 Corresponding author. of these ZFPs are confirmed at the protein level whereby E-MAIL [email protected]; FAX (734) 622-5668. Article and publication are at http://www.genesdev.org/cgi/doi/10.1101/ PPAR␥2 expression is virtually knocked out in the pres- gad.953802. ence of ZFP54, whereas only a low but detectable level of GENES & DEVELOPMENT 16:27–32 © 2002 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/02 $5.00; www.genesdev.org 27 Downloaded from genesdev.cshlp.org on September 23, 2021 - Published by Cold Spring Harbor Laboratory Press Ren et al. Figure 1. Structure of PPAR␥ gene and location of accessible chromatin in P2 promoter. (A) Genomic structure of mPPAR␥ showing exon splicing (Zhu et al. 1995). (B) Location of DNase I hypersensitive sites (DHS1 and DHS2) in the proximal P2 pro- moter, along with positions of ZFPs binding sites. (C) ZFP target sequences and finger designs. (Target sequence) The promoter DNA sequence to which each ZFP was designed. (Finger de- signs) The residues in each position from −1 to +6 of the recog- nition helix of each finger targeted against the cognate basepair ␥ triplet subsite. Figure 2. Ectopic expression of PPAR 2-specific ZFPs in 3T3- L1 cells. 3T3-L1 cells were infected with retroviral vectors ex- pressing PPAR␥2 promoter-specific zinc finger repressors, PPAR␥1 remains (Fig. 2D). In the presence of ZFP55 only ZFP54 and ZFP55. The uninfected 3T3-L1 cells (wild type) or PPAR␥2 expression is reduced substantially by ZFP55. cells infected with retroviral vector expressing LacZ gene were Because P1 and P2 are >63 kb apart, it was predicted that used as negative controls. (A) Total RNA was isolated from ZFPs targeted to P2 would selectively inhibit PPAR␥2 selected cell lines and determined the levels of ZFP mRNA by expression. This was the case with ZFP55, which gave Northern blot analysis using a 300 bp C-terminal fragment of 50% knockdown of PPAR␥2 only, however ZFP54 sup- ZFP as probe. The equivalence of RNA loading was verified by pressed both isoforms almost completely. ZFP54 and -actin. (B) Total RNA was isolated on indicated days postdif- ZFP55 bind to opposite strands of the DNA and have ferentiation and subjected to RNase protection assay using a slightly different DNA binding affinities, which may PPAR␥-specific riboprobe. Arrows indicate the protected contribute to their differential effects on P2. PPAR␥ and -actin mRNAs. The fold-induction was calculated The promoter specificity exhibited by the two ZFPs by normalization of PPAR␥-specific signal with -actin signal facilitates examination of the functional requirements and presented as fold-change compared to day-0 of wild-type for each PPAR␥ isoform in adipogenesis. In complete cells. (C) The real time quantitative RT–PCR (TaqMan) analysis concordance with inhibited mRNA and protein expres- of PPAR␥1 and PPAR␥2 mRNA expression. 40 ng and 60 ng of sion, we find that ZFP54 totally blocks the cellular lipid reversed transcribed total RNA was used for PPAR␥2 and accumulation that is a marker of adipogenesis (Fig. 3). PPAR␥1 analysis, respectively. (D) Western blot analysis with a Furthermore, the cells in which only PPAR␥2 is selec- polyclonal antibody against PPAR␥. The arrows indicate tively repressed by 50% (ZFP55) also show a correspond- PPAR␥1 and PPAR␥2 proteins. 28 GENES & DEVELOPMENT Downloaded from genesdev.cshlp.org on September 23, 2021 - Published by Cold Spring Harbor Laboratory Press Functional differences between PPAR␥1 and PPAR␥2 also was not inhibited, but rather an increase in its ex- pression was observed in the presence of ZFP54. This suggests a potential negative feedback mechanism whereby PPAR␥ may actively down-regulate C/EBP␦,in keeping with the reduction in C/EBP␦ observed during progression of adipogenesis (Wu et al. 1998). Taken to- gether these results support the notion that the inhibi- tory effects of ZFPs are specific. Genes that are expressed during adipogenesis after PPAR␥ induction include C/EBP␣ and the adipocyte-spe- cific aP2 (Wu et al.